The logical status of definite descriptions

Abstract

Ever since Bertrand Russell published his theory of descriptions in 1905, a debate about the proper logical and linguistic analysis of definite descriptions has ensued. In this essay I expound the theories of several linguistic philosophers who have been influential in shaping this debate, after which I present my own opinion on the issue. In the first chapter I discuss Russell's theory of descriptions and Strawson's reply to it. Russell argued that definite descriptions are quantified phrases and not referring expressions, and that, on such an analysis, certain logical puzzles which arise within accounts given by other philosophers could be solved. Subsequently, Peter Strawson claimed that Russell was mistaken: definite descriptions do not function as quantifiers but as referring expressions. The second chapter presents the second stage in the debate which was ushered in by the work of Keith Donnellan and Saul Kripke. Donnellan observed that definite descriptions have both an attributive and a referential use. Although there is general agreement that descriptions have these two uses, there is no agreement that they have two meanings. Many have appealed to ideas present in the work of Paul Grice to deny that descriptions also have a referential meaning. They argue that the referential use of descriptions is not semantically significant; it is merely a pragmatic phenomenon and, as such, it does not refute Russell's theory. Saul Kripke is perhaps the leading exponent of this line of thought. In the final chapter I expound the arguments of two contemporary theorists active in this field: Michael Devitt and Kent Bach. Devitt argues that definite descriptions have referential meanings, while Bach follows Kripke in holding that Donnellan's distinction boils down to a pragmatic one. In the conclusion I argue that while I believe that in ordinary language definite descriptions do have referential meanings, what we really need to be looking for is a definition of definite descriptions that is suited for the sciences - by sciences I mean not only the natural sciences but also the cultural sciences starting with philosophy. That is, we need to be looking for a way of analyzing sentences containing definite descriptions that makes their logical form explicit, one upon which we all agree and which would thus serve to make our scientific discourse more precise. And I claim that the solution should be based on Russell's analysis.